Gy2412371242

Suresh Palaka, Dr. K. Srinivasa Rao / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 2, Issue 4, July-August 2012, pp.1237-1242
Fault Tolerant Logic Design For Multiple Faults In
Combinational Circuits
SURESH PALAKA 1, Dr. K. SRINIVASA RAO 2
1 Research Scholar, JNTUH, Hyderabad , Andhra Pradesh
2 Principal & Professor, Department of ECE, TRRCE, Hyderabad

Abstract:-
A fault detection and analysis of stuck-at faults in synchronous sequential circuits. In this paper, the
multiple fault situation is developed in a analysis discussed in [3] is reviewed and slightly
combinational circuit with basic Boolean logic modified to present a new method for deriving the
gates. Especially Boolean logic gate is proposed for shortest test sequence. Treating the present and next
the computation of multiple faults in parallel bus state of the circuit as pseudo-input and pseudo-output
lines. During programming FPGA devices there is vector respectively, a vector Boolean
a probability of lines getting fused to either stuck difference technique is utilized to determine the set of
at zero or stuck at one faults. Where conventional input/state pair that will produce a difference in
diagnosing tools analyze stuck zero and stuck one output between fault-free and faulty
faults only multiple fault cases are not considered. circuits. Assuming that fault-free and faulty circuits
Using eminent logic fault tolerant design problem start in the same initial state, they must be driven by
is solved out in this paper. applying a sequence of length one input vectors to a
state in which a difference in output is evidenced. If a
Keywords: Multiple Fault, FPGA, Boolean logic, difference in output cannot be achieved immediately
stuck-at fault based on the difference in next state, the shortest
sequence, of length > 1, of input vectors is
I. INTRODUCTION determined, which when applied will propagate the
Fault analysis is currently one of the fault to produce an output difference.
principal research areas of digital systems theory.
One of the most important problems in this area is the II. DEFINITIONS AND NOTATION
generation of fault-detection test sets for Consider a Boolean function f(u1,…,un) of
combinational logic circuits. Much effort on this the Boolean variables u1,…un. Given a combinational
problem has resulted in some significant results logic circuit realizing this Boolean function, the lines
pertaining to problems characterized by the of the circuit is labeled using the terminal numbering
assumption that only single-fault situations can convention [3] where each line of the circuit is
occur in the circuits under consideration. Recently, labeled with an integer such that the output of any
however, there has been increasing emphasis given to gate is labeled with an integer greater than that used
multiple fault situations. Arguments for the to label the input terminals of the gate. In addition,
importance of this work have been made quite well the primary inputs will be denoted as u1,…,un. The
elsewhere and will not be repeated here except to say logical value of line j, j = 1,…,m, of the circuit
that it is evident that there are many situations in clearly depends on the values of the primary
which the possibility of more than one fault existing inputs u1,…,un, and is been denoted by this relation
in the system cannot be ignored. for all lines other than the primary output line as
Xj(u1,…,un). The logical value of the primary output
Fault detection in digital circuits has emerged as an line of the circuit can, of course, be expressed entirely
important principal research area in fault-tolerant in terms of the primary inputs, but it can also be
computing. The increasing complexity and popularity expressed in terms of the primary inputs and some
of today’s LSI realization of digital circuits have specified subset of lines of the circuit, say lines i1, i2,..,
rendered the problem of fault detection, fault and ip. This is denoted as F (u1,…,un, Xi1,…,Xip,). For
analysis, and test generation extremely difficult. example, consider the combinational logic circuit of
Significant results have been achieved in generating Fig. 1. The primary output can be expressed in terms
tests for combinational logic circuits under of the primary inputs as
the assumption that only the occurrence of single
faults of a stuck-at nature is more probable. However, .
there are cases where multiple fault situations must be
considered. Several authors [l], [2] have extended the However, the primary output based on 11 and 12 can
Boolean difference technique to the multiple fault be expressed as
case in combinational circuits, and [3] in

1237 | P a g e


Each line of the circuit is labeled with a unique
integer. A line numbered j has a logical value
which is dependent on the current
values of the input and state vectors. Assuming that
the initial state of the circuit is known as , the
behavior of the circuit can be described by vector-
valued Boolean output and next state functions as
follows:

(1)
Figure. 1:

(2)
In the following, to simplify the formulas which will where has the form
be developed, we will abbreviate this expression for
the logical value of the primary output line as
F(Xi2,...Xip). It will be understood that primary input
variables may also be present in the (3)
expression. Referring to the circuit of Figure 1, Iji,i = 1,2, . . . , p , is either primary input variable
from the set Xl,X2,. . . ,Xm, or one of the specified set
of internal variables of the circuit.
When it is clear that we are concerned with some In considering multiple faults, we must determine the
specific subset of the lines, say i1,…,ip, and have dominating faults and discuss about equivalent faults.
expressed the primary output accordingly as Two faults α, and β in the circuit starting in an initial
F(Xi2,...,Xip,), the expression resulting from setting state S’(0) are said to be strongly equivalent if the
Xi, = ai2,…,Xip = aip, where ai Ε {0,1}, j = 1,…,p, is output and next-state function of the circuit with fault
denoted as F(Xi, = ail ,Xi, = ai,) = F(ai1,…,ai,). α, are identical to those of the circuit with fault
Referring again to the circuit of Figure1, β. Strong equivalence of faults in a synchronous
sequential circuit reset-able to an initial state S’(0) is
or if it were clear that we were considering identical to the concept of equivalence of faults in the
F(X11,X12), it can be written as combinational circuit derived from the
synchronous sequential circuit by breaking all
feedback loops, and considering the current or present
III CIRCUIT AND FAULT DESCRIPTIONS state vector as a pseudo-input and the next-
Consider a synchronous sequential circuit state vector as a pseudo-output.
with m inputs, n outputs, and b-bits of memory as
shown in Figure 2. In vector notation, the circuit can IV BOOLEAN DIFFERENCE FORMULAS
be described as: FOR MULTIPLE FAULT ANALYSIS
Input Vector = (k) = [xi(k),…, xm(k)] The Boolean difference [1], [2] of a Boolean
Output Vector = (k) = [Y1(k),…,Yn(k)] function F(ul,….u,Xk) = F(Xk) with respect to the
State Vector = (k) = [Si(k), …, Sn( k )] variable Xk is defined as:

(4)
where 0 denotes the EXCLUSIVE-OR operator.
Similarly, the double Boolean difference of
F(ul,u2,…,UmXi,Xi) = F (Xi,Xj) with respect to Xi and
Xj is defined as:

(5)
Or
Figure 2: Synchronous Sequential Circuit

where k is the time parameter, and Xi;(k), 1 ≤ i ≤ m, (6)
Yi(k), 1 ≤ I ≤ n , and Si(k), 1 ≤ i ≤ b are Boolean
variables. And still more generally,
1238 | P a g e

The state vector S’ in Eqn. (10) must be replaced by
the initial state vector S’(0) as follows:

(7)

where F-F(Xi1,…,Xi) = F(ul,…,un,Xi1,…,Xip).
(11)
Some other interesting properties of the Boolean
difference and the EXCLUSIVE-OR operation which
However, if the output functions do not depend on the
will be useful in the following are:
selected internal variables with the initial state
setting, i.e., Eqn. (11) = 0, the vector Boolean
difference of the next state
. with respect to should be considered as in the
next equation:
(8)

and

(12)
(9)
The vector Boolean differences in Eq.(11) and (12)
V. BOOLEAN DIFFERENCE EXPRESSION equal to one if any term in the sum equals to one, and
FOR MULTIPLE FAULTS IN SEQUENTIAL equal to zero only if every term in the sum equals to
CIRCUITS zero. Our next objective is to solve Eq. (11) for the
input/state pairs that will sensitize the output to the
Consider a fault involving p internal variables I’= [I j1,
differences in the specified internal variables.
Ij2,…,Ijp] simultaneously. The internal inputs can be
However, if the output functions do not depend on the
primary inputs, primary outputs, and memory lines.
selected internal variables, then Eq. (12) must be
The vector Boolean difference of the output Y’ with
solved for the input/state pairs that sensitize the next
respect to I’ is defined as [3].
state to the differences in the internal variables under
consideration. The simplified forms for Eq.(11) and
(12) are [2]:
(10)

One of the modifications to Goldstein’s work [3]
should be pointed out here. As mentioned earlier,
since Figure. 3 is just a conventional combinational (13)
circuit, Eqn. (10) should be modified with a sense of
sequential circuit, i.e., the state vectors cannot be
treated as primary input as in a “pure” combinational
circuit.
(14)

Where, 1Gα = 1G[X’,S’(0), p-tuple of decimal
equivalent of α], 1Fα = 1F[X’,S’,p-tuple of decimal
equivalent of α], and mr = p-tuple of decimal
equivalent of r.

VI DETECTING STAGES
The circuit is in a detecting state for a fault
a, if an applied input can produce a difference in
output between fault-free and fault-n circuits.
Suppose that at time t the fault-free and fault-a
circuits are in state Sff ( t ) and Sπ(t), respectively. Let
Figure 3. Combinational Circuit obtained from Sff(t) = Sx and Sπ (t) = Sy where Sx may be the same as
Synchronous Sequential Circuit of Figure 2. Sy. The pair (Sx,Sy) is called a detecting pair of states,
or shortly detecting pair, if there exists at least one
input vector which when applied to both circuits will

1239 | P a g e

produce a difference in output between them at time VIII APPLICATIONS AND DISCUSSION
(t + 1) which means that the fault π is detected. In the previous section we have developed
Boolean difference formulas for various types of
Goldstein [3] has used a successor tree for keeping multiple fault situations. These results are not only
track of the states reached and the outputs generated interesting in the sense that they give a physical
by the fault-free and faulty circuits. We use a meaning to expressions which have heretofore had no
different technique called detecting tree in which the such interpretation associated with them, but are also
transition tables of both circuits are used to find the clearly useful in the fault analysis of combinational
set of detecting pairs Σ*π for fault π, and to form the circuits. The attractive completeness of the Boolean
detecting tree searching for the shortest input difference concept has been extended to the multiple
sequence which will lead the fault-π circuit from the fault case. For example, suppose we are given some
initial state to a detecting state. multiple fault. To use the developed results, we
would first determine the corresponding reduced
Since faulty lines may be memory or state lines, and multiple fault, and then express the Boolean function
since we are looking for test sequences of length realized by the circuit in terms of the primary inputs
greater than one, we may drive some of the faulty and the lines on which the reduced multiple fault
state lines to logical values that are complements of existed. If two or more components of this reduced
their faulty values in the process of driving the fault-a multiple fault were on paths of the circuit which
circuit through some different states before reaching emanated from the same fan-out point, or if one of
the first detecting state. So for simplicity, we look for the components was at a fan-out point and one or
both type I and type I1 test sets for fault π. more of the remaining components were on paths
emanating from this fan-out point, then with the
resulting Boolean expression, we would be
VII DETECTING TREE effectively considering an equivalent circuit where
If the test sequence of length one does not such a point of fan-out has been removed. This is a
exist, or the output difference between the two significant observation since it indicates that the
circuits cannot be obtained immediately, we can try results of the previous section involve an algebraic
to find a test sequence of length greater than one, if form of the conversion of a circuit to an equivalent
one exists, from the detecting tree. The procedure for fan-out-free circuit used by Schertz and Metze [4] for
construction of the detecting tree is as follows: the purposes of making combinational circuits readily
diagnosable for multiple faults.
Step 1: Form the output and next-state equations and
solve Eqns.(9) and (10). Case I (Eqn. (9) = 0, Eqn.
(10) # 0): This indicates that the output functions do IX RESULTS AND OBSERVATIONS
not depend on the specified multiple fault and the
selected internal variables at all, but the next state
functions do.
Case I1 (Eqn. (9) # 0): This is the case in which the
output functions depend on the specified multiple
fault, but not on the selected internal variables. Note
that there is no need to check if Eqn.(lO) # 0 in this
case, Case I11 (Eqns.(S) and (10) = 0): This indicates
that the fault is undetectable.

(15)
Figure 4. Write operation

+…
+ +…
+ …..
(16)

1240 | P a g e


Figure 8: RTL view of the implemented system using
Xilinx synthesizer.

Figure 5. Read Operation X CONCLUSION
This paper contributes in developing
simplified Boolean logic architecture for the detection
and analysis of multiple faults in programmable
devices. A diagnosing approach for multiple faults at
one time in various lines or locations is been
suggested the simulation observations were carried
out for the suggested approach defined using VHDL
definition. The fault detecting system is analyzed for
both sequential and combinational circuits with stuck-
at one and stuck-at zero fault simultaneously.

XI REFERENCES
[1] C. T. Ku and G. M Mason, “The Boolean
Differ ence and Multiple Fault Analysis”,
Figure 6. Summarized synthesis report for the ZEEE Trans. on Computers, Vol. C-21,
developed estimation system. No.1 Jan 1975, pp.62-71.
[2] S. R. Das, P. K. Srimani, and C R. Dutta,
“On Multiple Fault Analysis in
Combinational circuits by Means of Boolean
The obtained power analysis report is obtained as, Difference”, P.m of the ZEEE Vol 64, No.9,
Sept.1976, pp.1447-1449
[3] L.H Goldstien “Analysis of Multiple Faults
in Synchronous Sequential Circuits by
Boolean Difference Techniques”, Private
Communication
[4] D R Schertz and G Metze “On the design of
multiple fault diagnosable network”, IEEE
Trans. Comput., 1361-1364, Nov 1971.
vol.C-20,pp
[5] A Dehon and M. J Wilson, “Nanowire based
Sub lithographic programmable logic
arrays”, in Proc. Int. Symp, Field program
Gate Arrays, Feb 2004, pp 123-132
[6] H. Naeimi and A DeHon, “Fault secure
encoder and decoder for memory
Figure 7. Power analysis applications”, in Proc. IEEE Int Symp
Defect Fault Tolerance VLSI Syst.,
Timing report for the implementation is observed as, Sep 2007, pp.409-417
[7] S.J. Piestrak, A.Dandache and F Monterio,
“Design ing fault secure parallel encoders
for systematic linear error correcting
codes”, IEEE Trans. reliability , vol.52
no.4, pp.492-500, jul 2003.

1241 | P a g e

[8] Naeimi H and DeHon A, “Fault-tolerant
nano-memory with fault secure encoder and
decoder,” presented at the Int. Conf. Nano-
Netw., Catania, Sicily, Italy, Sep. 2007.

About the authors:

Mr. Suresh Palaka is working as an
Assistant Professor in the Dept of ECE,
HITS, Hyderabad, AP, India He is
currently pursuing Ph. D. from JNT
University, Hyderabad, India. His
research interests include Fault tolerance, Testability,
VLSI and Embedded Systems.

Dr. K. Srinivasa Rao received Ph.. D.
degree from Andhra University, India.
He is currently working as Professor of
ECE and Principal, TRRCE, Hyderabad,
AP. He has 27 years of experience in the
field of education. He is an active
member of IEEE, ISTE, SEMCEI, BMESI. He is a
Fellow of IE(l), and IETE. He has several
publications in various journal and conferences of
National and International repute. He is a resource
person for SONET, Visiting professor for ISTE .
He organized many technical events. His research
interests include Communications, VLSI and
Embedded Systems. He is guiding 09 scholars
affiliated to JNTUH & on these submitted to JNTUH.
Awaiting for vivo.

1242 | P a g e

Gy2412371242

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (20)

Similar to Gy2412371242

Similar to Gy2412371242 (20)

Recently uploaded

Recently uploaded (20)

Gy2412371242